Views
No views yet
| Model | Download | Download (with sample test data) | ONNX version | Opset version | Top-1 accuracy (%) |
|---|---|---|---|---|---|
| Inception-1 | 28 MB | 29 MB | 1.1 | 3 | |
| Inception-1 | 28 MB | 29 MB | 1.1.2 | 6 | |
| Inception-1 | 28 MB | 29 MB | 1.2 | 7 | |
| Inception-1 | 28 MB | 29 MB | 1.3 | 8 | |
| Inception-1 | 28 MB | 29 MB | 1.4 | 9 | |
| Inception-1 | 27 MB | 25 MB | 1.9 | 12 | 67.23 |
| Inception-1-int8 | 10 MB | 9 MB | 1.9 | 12 | 67.24 |
| Inception-1-qdq | 7 MB | 5 MB | 1.12 | 12 | 67.21 |
Compared with the fp32 Inception-1, int8 Inception-1's Top-1 accuracy drop ratio is -0.01% and performance improvement is 1.26x.NoteThe performance depends on the test hardware. Performance data here is collected with Intel® Xeon® Platinum 8280 Processor, 1s 4c per instance, CentOS Linux 8.3, data batch size is 1.
data_0: float[1, 3, 224, 224]prob_1: float[1, 1000]wget https://github.com/onnx/models/raw/main/vision/classification/inception_and_googlenet/inception_v1/model/inception-v1-12.onnx1bash run_tuning.sh --input_model=path/to/model \ # model path as *.onnx
2--config=inception_v1.yaml \
3--data_path=/path/to/imagenet \
4--label_path=/path/to/imagenet/label \
5--output_model=path/to/save